Dozens of epidemiologic reports have linked insomnia to increased risk for suicidal ideation, suicidal behavior and suicide death in patients with major depression. The mechanism whereby insomnia increases the intensity of suicidal ideation may be mediated through dysfunctional beliefs and attitudes about sleep, somewhat similar to hopelessness. We have unpublished, preliminary data showing that the addition of the hypnotic eszopiclone to open-label fluoxetine in the treatment of depressed insomniacs is associated with a reduction in suicidal ideation, as compared with placebo added to fluoxetine. We now propose to confirm the premise that treatment of insomnia reduces suicidal ideation in a multi-site clinical trial. Wake Forest University (WFU) will be the coordinating site and a recruitment site, while Duke University (DU) and University of Wisconsin (UW) will also serve as recruiting sites. Outpatients (N=138) with major depression, insomnia, and mild- moderate suicidal ideation will be treated with open label fluoxetine for 8 weeks and will be further randomized to receive either eszopiclone or placebo at bedtime for the same period. Patients will have return office visits at 1, 2, 4, 6, and 8 weeks after treatment initiation. Assessments wll include measures of suicidal ideation intensity, overall depression severity, insomnia severity, dysfunctional cognitions about sleep, nightmare intensity, hopelessness, and actigraphy. All data will be entered in a WF-created web-based interface, and consistency of methodology across sites will be assured with regular teleconferences between sites. The sample size will be sufficient to allow 80% power to detect a 2.0 point difference in the Beck Scale for Suicide Ideation between treatment arms. Safety of participants will be assured by (1) exclusion of patients with more than severe suicidal ideation at baseline, (2) frequent follow up, (3) limited access to hypnotics, (4) access to university psychiatric inpatients units for psychiatric emergencies, (5) 24-hour per day emergency services available through the sites respective psychiatric residency programs, (6) involvement of families and loved ones, when available, in the consent process, (7) and the creation of a Data Safety Monitoring Committee that will include expertise in depression, insomnia, clinical trials, statistics, suicidology, biostatistics, ethics, and a patient advocate. The primary aim will be to assess the effect of treating insomnia with hypnotic medication on the intensity of suicidal ideation. The secondary aim will be to test whether reductions in suicidal ideation in depressed insomniacs is mediated through either reduced dysfunctional beliefs about sleep, reduced hopelessness, or through reduction in nightmares. The results of this study will inform the appropriate management of patients with mild-moderate suicidal ideation, insomnia and depression - representing an extremely common clinical scenario.